Better C

rops/Vol. 100 (2016, No. 2)

19

Abbreviations and notes: N = nitrogen; P = phosphorus; K = potassium; S = sulfur; Ca = calcium; Mg = magnesium; Fe = iron; Zn = zinc; Cu = copper. IPNI Project SEAP-05

SOUTHEAST ASIA

Research documented nutrient removal from harvested oil palm fresh fruit bunches (FFB). Reports from South-east Asia and Africa show that each t of FFB contains

3.0 to 5.0 kg N, 0.3 to 0.7 kg P, 3.5 to 5.3 kg K, and 0.5 to 0.9 kg Mg (Tinker and Smilde, 1963; Ng and Thamboo, 1967; Tarmizi and Mohd Tayeb, 2006; Prabowo et al., 2006; Donough et al., 2014). There are reports of nutrient contents in post-milling residues after processing and extraction of palm oil and palm kernels from the FFB. These are for the empty fruit bunches (EFB) and the palm oil mill effluent (POME). The nutrients in the post-milling residue do not fully reconcile with pre-milling values (Prabowo et al., 2006). This suggests that the palm oil and palm kernel may contain some of them, or there could be unaccounted loss in the milling process. In a review of the fate of plant nutrients in palm oil production, Corley (2009) wrote that, “palm oil contains no N or K, and only about 20 g of P per tonne.” Beyond this, there is virtually no information for contents of plant nutrients in palm oil.

The International Plant Nutrition Institute (IPNI), received an inquiry about the S content in palm oil in 2014. At the time, the Southeast Asia Program (SEAP) of IPNI sampled FFB in a project in Kalimantan for bunch analysis (BA, procedure shown in Figure 1, from Oberthür et al., 2012) to estimate yield of crude palm oil (CPO) and palm kernel. For this paper, the ex-tracted samples of palm oil were analyzed. First, to determine the content of plant nutrients in CPO from BA (referred to as ‘BA CPO’). This will show the portion of exported nutrients in FFB that the oil contains. Second, since the Kalimantan project compared different fertilizer managements, the results should also indicate if these had any influence on nutrient contents in the BA CPO. Third, to compare BA CPO with CPO from an industrial mill to make a balance sheet to identify where nutrient losses occur in the palm oil milling process.

The Kalimantan Project The Kalimantan project validated the hypothesis that

applying fertilizer more frequently, in line with 4R Nutrient Stewardship, will improve nutrient use efficiency (NUE) on sandy soils, and increase yields. In our project ’nutrient best management practice’ (NBMP), the application of fertilizer mixtures supplying N, P, K, Mg, S, and B four times yearly, was compared to standard estate practice (SEP) where single nutrient fertilizers are applied once or twice annually (Table 1). The project included a reduced fertilizer rate treatment with 80% of the full rate. There were four treatments in a fac-torial design of two application frequencies (NBMP and SEP)

and two application rates (full and reduced). There were three replicates: full-sized blocks, each 25 ha. Within each block, we embedded two plots each of 36 palms (of which the central 16 palms were recorded). One was fertilized in the same way as the rest of the block, the other left unfertilized.

Bunch Analysis versus Palm Oil Mill ProcessingThe BA CPO is obtained directly from individual FFBs

(Figure 1), whereas the palm oil mill CPO is extracted from large batches of FFB (Figure 2). In the BA process, bunches are processed ‘fresh’, whereas the FFB in the mill is sterilized (cooked) using steam under pressure. BA bunches are chopped to separate the stalk from fruit-bearing spikelets, which are then sprayed with ethrel® (active ingredient is ethephon or 2-chloroethylphosphonic acid, C

2H

6ClO

3P, 21% P) to accel-

erate fruit abscission, and the fruits are separated from the

By Christopher R. Donough, Angger Cahyo, Ruli Wandri, Myles Fisher, and Thomas Oberthür

Plant Nutrients in Palm Oil

Figure 1. Bunch analysis procedure (Oberthür et al., 2012) imple-mented by IPNI Southeast Asia Program in Kalimantan, Indonesia.

An apparent knowledge gap concerning the amount of plant nutrients in palm oil motivated a study to determine plant nutrient content in palm oil and assess the impact of fertilizer management on such content.

Export of plant nutrients was low in palm oil extracted by industrial mills; part of the nutrients likely remain in post-milling residues.

Selected nutrients in palm oil were affected by fertilizer application rate, but not timing or frequency.

Extraction (16-20 hr)

Fresh Fruit Bunch

Chopping

Spikelets

Detaching

Fruits

Depericarping

Mesocarp Nuts

Oven Drying

Dry Mesocarp Dry Nuts

Mincing & Sieving Cracking

Kernels

Fibre

Oil

Evaporation

Hexane

Hexane

HexaneOil +

Ethephon

Stalk

Empty Spikelets

Shell

20

Bet

ter

Cro

ps/V

ol. 1

00 (

2016

, No.

2)

spikelets manually. Each BA sample of 4 to 6 kg is sprayed with 250 ml of a 0.05% v/v solution so the concentration of P applied in ethephon per sample is 4 to 6 mg/kg. A sample will contain (approximately) the following components (expressed as % dry matter): spikelets (13%) and fruits, which consist of mesocarp (69%; containing oil that approximates 75 to 80% of the dry mesocarp), kernel (10%) and shell (8%). Distribution of the P from ethephon between the sample components is

not known, but not likely all of the 4 to 6 mg/kg added P will be in the extracted oil. In the palm oil mill, the cooked FFB is fed into a rotating drum thresher that strips fruits from the bunches (Figure 2).

In the BA process, the fruits are then manually separated into the mesocarp (that contains the CPO) and the seed nuts (that contain the palm kernels). The BA CPO is extracted from a sample of the (oven-dried and sieved) mesocarp using a soxhlet extractor with hexane (C

6H

14) as the solvent. In the

palm oil mill, after cooking the fruits are fed into a rotating drum stripper and then into a digester where further heating with steam loosens the mesocarp from the nuts. Stirring arms in the digester help the process. The fruit ‘digest’ is then fed into a screw press where oil is pressed from the mixture of water, mesocarp fibres and nuts. The oil from the press (‘crude oil’ in Figure 2), is the ‘initial crush’ CPO, which is still mixed with solids from the ‘digest’. The ‘initial crush’ CPO is screened to remove the larger non-oil particles, and then goes to a

Figure 2. Typical milling process in a palm oil mill receiving fresh fruit bunches (FFB, top of figure) and extracting crude palm oil (CPO, left middle) and palm kernels (PK, right middle). From Siew (2011). Numbers shown for various components are from original author and do not add up.

Fresh Fruit Bunches (1000kg)

Sterilization

Stripping

Digestion

Pressing

Screening

Clarification

Desanding Purification

Separation Drying

Nut/FibreSeparation

Nut Drying

Nut Cracking

Dry Separation

Wet Separation

Kernel Drying

STEAM CONDENSATE

EMPTY BUNCHES (234kg)

FIBROUS MATERIALS FRUITS (666kg)

STEAM

WATER CRUDE OIL

OIL

DIRT

SLUDGE(180kg)

CRUDE PALM OIL(225kg)

WASTEWATER

PRESS CAKE

FIBRE (180kg)

CRACKED MIXED

DRY SHELL(total shell

73kg)

WET SHELL

FUELKERNELS(67kg)

Table 1. Nutrient Management Practices in IPNI SEAP Kalimantan Project.

Treatment - - - - - - - - - - - - - NBMP1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - SEP2 - - - - - - - - - - - - - - - - - - - - - - - - - -Nutrient sources3 Urea Gr Ammophos Kornkali+B Urea Pr TSP MOP Kieserite Borate N-P-K-Mg-S-B contents 46-0-0-0-0-0 16-9-0-0-12-0 0-0-33-4-4-0.8 46-0-0-0-0-0 0-46-0-0-2-0 0-0-50-0-0-0 0-0-0-16-22-0 0-0-0-0-0-15 Annual frequency of application 4 4 4 2 1 2 1 1Method of application Mixed and applied manually Each type individually applied manually

Placement Broadcast outside palm circles onto heaps of pruned fronds

Urea and borate applied onto soil surface inside palm circles; other fertilizers broadcast outside palm circles onto heaps of pruned fronds

Approximate4 annual application rate, kg/ha - Full5 rates shown

Nitrogen, N 150 150 Phosphorus, P 13 12 Potassium, K 230 230Magnesium, Mg 25 26Sulfur, S 28 <1 34Boron, B 2 11Nutrient Best Management Practice; 2Standard Estate Practice; 3Urea Gr = granular urea, Urea Pr = prilled urea, TSP = triple superphosphate, MOP = muriate of potash or KCl; 4Values vary slightly (less than +/- 5%) year-to-year; 5Full rate based on annual FFB yield target = 21 t/ha, calculated for NPKMg only, S rate varies with type of fertilizers used; Reduced rate is approx. 80% of Full rate.

Oil palm mill receiving fresh fruit bunches for processing, Kalimantan, Indonesia.

IPN

I/C.D

onou

gh

Better C

rops/Vol. 100 (2016, No. 2)

21

clarification tank where much of the remaining non-oil solids settle. The cleaner CPO is skimmed off the top and passes through a centrifuge to remove remaining solids, followed by drying under vacuum. The cleaned crude palm oil passes to a production tank as ‘Mill CPO’, which then goes on to palm oil refineries. Samples of BA CPO and Mill CPO were obtained for analysis and comparison.

Nutrients in Mill Crude Palm Oil and Bunch Analysis Crude Palm Oil

Samples from the palm oil mill production tank are repre-sentative of CPO sold to refineries. The analyses show that there are only very small amounts of plant nutrients exported with the Mill CPO (Table 2). The BA CPO values were many times higher than the mill CPO values for all nutrients except Cu (Table 2). The difference is likely attributable to differences in the oil extraction methods at the mill (mechanical pressing) and in BA (solvent extraction). An important difference is that in the palm oil mill, FFBs are pressure cooked prior to press-ing, so that nutrients may be lost dissolved in the sterilizer condensate (Figure 2). In BA, FFBs were not pre-treated.

Further loss of nutrients may occur after pressing in the palm oil mill when the oil is clarified and cleaned to remove dirt and impurities, and dried. Indeed, the nutrients contained in initial crush CPO from the mill were even higher than the BA CPO (Table 3). This is likely because the BA process used clean samples of fruit mesocarp and the leachate in soxhlet process contains only solutes. In contrast, the CPO after crushing in the palm oil mill contains much solid materials from the fruit digest.

Influence of Fertilizer Management Reducing the fertilizer applied to 80% of the

full rate affected N and Ca in BA CPO, which increased (Table 4). BA CPO from unfertilized plots showed significantly lower contents of P, K, and Mg compared to plots that received the same fertilizers as the blocks (Table 5). This suggests that the additional P, K, and Mg in the CPO came from added fertilizers. Other nutrients in the fertilizers did not affect contents of the oil.

The results for P could be slightly increased by 4 to 5% by P added via the ethephon treat-ment, which added an estimated 4 to 6 g/t to the BA samples.

Removal Rate from PlantationsContents of N, P, K, and Mg in FFB from the

Kalimantan project had been reported earlier

(Donough et al., 2014). Those results are reproduced in Table 6, with addition of results for Ca and S previously not reported. Each t of FFB removed approximately 3 kg N, 0.4 kg P, 3.8 kg K, 0.6 kg Mg, 0.5 kg Ca, and 0.3 kg S. Assuming a FFB yield of 25 t/ha, this translates to per ha removal of 75 kg N, 10 kg P, 95 kg K, 15 kg Mg, 12.5 kg Ca, and 7.5 kg S. The propor-tion of total N and K in FFB contained in the BA CPO is <1%, compared with almost 10% of P. In the case of S, about 3% is in the BA CPO. Nutrient removal (per ha) in the oil, assuming the same 25 t/ha FFB yield above, is just 0.6 kg N, 0.9 kg P, 0.1 kg K, 0.4 kg Mg, 0.3 kg Ca, and 0.2 kg S.

A sample for fresh fruit bunch analysis (main image). Fruit mesocarp is sepa-rated from each seed nut (bottom left) and then oil is extracted from the mesocarp by solvent extraction (bottom right).

IPN

I/C.D

onou

gh

Table 2. Nutrient contents in crude palm oil (CPO) from the mill and bunch analysis (BA), in g per tonne.

Nutrient N P K Mg Ca S Fe Zn CuMill CPO1 44 18 <10 3 9 11 3 <0.5 <0.5BA CPO2 93 145 20 58 42 37 55 3.0 0.51 Samples from palm oil mill (physical extraction by mechanical presses), taken from production oil tank after clarifying, cleaning and drying. Mean values of 3 samples.2 Samples from bunch analysis process (solvent extraction); mean values of 12 composite samples.

Table 3. Nutrient contents in crude palm oil (CPO) from different milling stages, g per tonne.

Nutrient N P K Mg Ca S Fe Zn CuMill CPO1 44 18 <10 3 9 11 3 <0.5 <0.5Initial crush CPO2 853 81 1,103 286 342 144 52 2.0 1.11 Samples from palm oil mill (physical extraction by mechanical presses), taken from pro-duction oil tank after clarifying, cleaning and drying. Mean values of 3 samples analysed by IMS.2 Samples from palm oil mill crude oil tank immediately after pressing, results are aver-age values from determinations by two laboratories (SRC and A&L), except for the N result from SRC only.

Table 4. Effect of nutrient management on nutrient contents in bunch analysis crude palm oil, in g per tonne.

Nutrient N P K Mg Ca S Fe Zn CuNBMP1 blocks 102 145 20 56 40 34 52 2.5 <0.5SEP2 blocks 85 145 20 60 44 39 58 2.9 <0.5Full-rate blocks 73 142 19 55 36 42 54 3.3 <0.5Reduced-rate3 blocks 113 148 21 61 48 32 56 2.6 <0.51 Nutrient best fertilizer management = nutrients mixed and applied 4 times a year (i.e., high frequency for every nutrient).2 Standard estate practice = nutrients applied singly, 1 to 2 times a year.3 80% of full rate.

22

Bet

ter

Cro

ps/V

ol. 1

00 (

2016

, No.

2)

ConclusionsThis work closes a significant gap in oil palm relevant

nutrient management knowledge, by clarifying the nutrient content in CPO.

CPO obtained from the palm oil milling process contains few plant nutrients (Table 2), indicating that most of the nu-trients in FFBs reaching the mills must remain somewhere in the mill system. Therefore, recycling nutrients from the mills back to plantations is an opportunity for better nutrient use efficiency in palm oil production. Decision models on alterna-tive uses for post-milling residues (e.g., for power generation) must factor in the opportunity cost of such nutrient recycling, as prices of fertilizers and fuel fluctuate.

Nutrient content in BA CPO was higher than that in Mill CPO (Table 1), most likely due to the difference between extraction by mechanical pressing of whole bunches and leaching with a solvent. While useful for monitoring at the experimental or plantation level, costly solvent extraction is not used at the industry level.

Fertilizer management, in this case different frequency of application and rates, had no effect on most plant nutrients in the extracted CPO. The contents of N and Ca fell when the rate of applied fertilizer increased from 80% to 100%. Explanations are uncertain without further investigation.

Applying fertilizers compared with no fertilizer increased P, K, and Mg in the CPO, indicating that some of the applied nutrients ended up in the oil component of FFB.

AcknowledgementsFunding for and coordination of all laboratory analyses of

oil samples in Canada was arranged by Dr. Jim Beaton and Dr. Eric Pedersen of Sulvaris Inc., Canada. This work was first inspired by Dr. Beaton’s timely query on S content in palm oil. N determinations were done at the University of Saskatchewan by Dr. Renato de Frietas. Palm oil samples from the IPNI SEAP project, as well as samples from the PT Sungai Rangit palm oil mill, were collected by research staff of PT Sungai Rangit (SR). The IPNI SEAP project was a collaboration with K+S Kali GmbH and PT SR (a subsidiary of PT Sampoerna Agro Tbk). BC

Mr. Donough is Senior Oil Palm Adviser to IPNI SEAP (E-mail: crdonough@gmail.com). Mr. Cahyo is Research Assistant and Mr. Wandri is Agronomy Research Manager at PT Sampoerna Agro Tbk. Dr. Fisher is a Consultant for IPNI SEAP, and Dr. Oberthür is Direc-tor, IPNI SEAP, Penang, Malaysia.

ReferencesCorley, R.H.V. 2009. The Planter 85(996):133-147.Donough, C.R., A. Cahyo, T. Oberthür, W. Ruli, J. Gerendas, and A.R. Gatot.

2014. Improving nutrient management of oil palms on sandy soils in Ka-limantan using the 4R concept of IPNI. Presented at International Oil Palm Conference, Bali, Indonesia.

Ng, S.K. and S. Thamboo. 1967. Malaysian Agricultural Journal 46:3-45.Oberthür, T., C.R. Donough, K. Indrasuara, T. Dolong, and G. Abdurrohim.

2012. Successful Intensification of oil palm plantations with best manage-ment practices: Impacts on fresh fruit bunch and oil yield. Presented at International Planters Conference, Kuala Lumpur, Malaysia.

Prabowo, N.E., H.L. Foster, and A.J. Silalahi. 2006. Recycling oil palm bunch nutrients. Presented at International Oil Palm Conference, Bali, Indonesia.

Siew, W.L. 2011. Palm oil: Edible oil processing. Accessed and printed from http://lipidlibrary.aocs.org/OilsFats/content.cfm?ItemNumber=40334 updated March 7th, 2011.

Tarmizi, A.M. and D. Mohd Tayeb. 2006. J. Oil Palm Res.18:204-209. Tinker, P.B.H. and K.W. Smilde. 1963. Plant and Soil 19:350-363.

Soil Test Levels in North America: 2015 Summary Update

The 2015 summary update booklet provides interpretive analysis of the results of IPNI’s most recent survey of soil test levels for North America. For special orders contact IPNI at circulation@ipni.net. Or visit https://store.ipni.net.

Table 5. Nutrient contents in bunch analysis crude palm oil - with and without fertilizers, in g per tonne.

Nutrient N P K Mg Ca S Fe Zn CuFertilized1 plots 79 142 22 61 45 33 55 2.6 0.5Unfertilized2 plots 77 128 17 50 45 35 57 2.6 0.51 Embedded plot in each block receiving the same fertilizer treatment as rest of the block. 2 Unfertilized plot in each block.

Table 6. Nutrient contents in fresh fruit bunches (FFB) and in bunch analysis crude palm oil (BA CPO), in kg per tonne.

Nutrient N P K Mg Ca SWhole FFB1 3.07 0.38 3.84 0.62 0.51 0.28Proportion in BA CPO2 0.02 0.04 0.01 0.01 0.01 0.01 Proportion in BA CPO in % 0.82 9.54 0.11 2.31 2.11 3.31 1 Contents in whole bunches, including CPO still in mesocarp. Mean of all four treatments.2 Contents in the oil extracted from whole bunches (bunch analysis CPO) and assuming an oil content of 25% in FFB.